study protocol open access study protocol for a randomized, … · 2017-08-25 · cognitive...

TRIALSKingston et al. Trials 2014, 15:72http://www.trialsjournal.com/content/15/1/72

STUDY PROTOCOL Open Access

Study protocol for a randomized, controlled,superiority trial comparing the clinical andcost- effectiveness of integrated online mentalhealth assessment-referral-care in pregnancy tousual prenatal care on prenatal and postnatalmental health and infant health and development:the Integrated Maternal Psychosocial Assessmentto Care Trial (IMPACT)Dawn Kingston1*, Marie-Paule Austin2, Kathy Hegadoren1, Sheila McDonald3, Gerri Lasiuk1, Sarah McDonald4,Maureen Heaman5, Anne Biringer6, Wendy Sword4, Rebecca Giallo7, Tejal Patel4, Marie Lane-Smith1

and Sander Veldhuyzen van Zanten1

Abstract

Background: Stress, depression, and anxiety affect 15 to 25% of pregnant women. However, fewer than 20% ofprenatal care providers assess and treat mental health problems and fewer than 20% of pregnant women seekmental healthcare. For those who seek treatment, the lack of health system integration and existing barriersfrequently prevent treatment access. Without treatment, poor prenatal mental health can persist for years andimpact future maternal, child, and family well-being.(Continued on next page)

* Correspondence: [email protected] of Alberta, 11405-87th Avenue, Edmonton T6G 1C9 AB, CanadaFull list of author information is available at the end of the article

© 2014 Kingston et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly credited.

mailto:[email protected]

http://creativecommons.org/licenses/by/2.0

Kingston et al. Trials 2014, 15:72 of 20http://www.trialsjournal.com/content/15/1/72

(Continued from previous page)

Methods/Design: The purpose of this randomized controlled trial is to evaluate the effectiveness of an integratedprocess of online psychosocial assessment, referral, and cognitive behavior therapy (CBT) for pregnant women comparedto usual prenatal care (no formal screening or specialized care). The primary outcome is self-reported prenatal depression,anxiety, and stress symptoms at 6 to 8 weeks postrandomization. Secondary outcomes are postpartum depression,anxiety, and stress symptoms; self-efficacy; mastery; self-esteem; sleep; relationship quality; coping; resilience; Apgar score;gestational age; birth weight; maternal-infant attachment; infant behavior and development; parenting stress/competence; and intervention cost-effectiveness, efficiency, feasibility, and acceptability. Pregnant women are eligible ifthey: 1) are <28 weeks gestation; 2) speak/read English; 3) are willing to complete email questionnaires; 4) have no,low, or moderate psychosocial risk on screening at recruitment; and 5) are eligible for CBT. A sample of 816 women willbe recruited from large, urban primary care clinics and allocation is by computer-generated randomization. Women inthe intervention group will complete an online psychosocial assessment, and those with mild or moderate depression,anxiety, or stress symptoms then complete six interactive cognitive behavior therapy modules. All women willcomplete email questionnaires at 6 to 8 weeks postrandomization and at 3, 6, and 12 months postpartum. Clinic-basedproviders and researchers conducting chart abstraction and analysis are blinded. Qualitative interviews with 8 to 10healthcare providers and 15 to 30 intervention group women will provide data on feasibility and acceptability of theintervention. Results of this trial will determine the feasibility and effectiveness of an integrated approach to prenatalmental healthcare and the use of highly accessible computer-based psychosocial assessment and CBT on maternal,infant, and family-based outcomes.

Trial registration: ClinicalTrials.gov Identifier: NCT01901796

Keywords: psychosocial assessment, online, screening, cognitive behavior therapy, pregnancy, depression, anxiety,stress, randomized controlled trial

BackgroundPrenatal mental health problemsDepression, anxiety, and stress are common in pregnancy.One in four pregnant women experiences symptoms ofdepression, stress, or anxiety, with 25% having mild tomoderate symptoms [1]. Without treatment, up to 48%of women with prenatal anxiety and 70% of those withprenatal depression [2] continue to experience symptomsthrough the postpartum period [3-5] and into their chil-dren’s early years of life [6-8]. The consequences of poorperinatal mental health are enduring. Two decades ofwell-conducted longitudinal studies demonstrate that evenmild to moderate perinatal distress can have serious ad-verse effects on mothers and children, including pretermbirth and low birth weight [9], child developmental delay[7,10,11], and poor child mental health [12,13].

The cycle of under detection and under treatment ofprenatal depression, anxiety, and stressTo date, perinatal mental healthcare has focused almostexclusively on preventing and treating postpartum depres-sion. This paradigm does not reflect current evidence that50 to 70% of postpartum anxiety and depression begin[14] and frequently co-occur [15-17] in pregnancy, nordoes it reflect the enduring effects of poor prenatal mentalhealth on child health [11,18,19]. Prenatal depression, anx-iety, and stress are severely under detected and under

treated, and two-thirds of women with substantial symp-toms remain unidentified by most obstetrical providers[20,21]. A number of barriers prevent women from seekingmental healthcare during the perinatal period, includingstigma, fear of being prescribed medication, lack of know-ledge about whether their symptoms are ‘normal’ or ‘abnor-mal’, and fear that their concerns will be dismissed [22-24].However, despite recommendations [25,26] and acceptanceby both healthcare providers [27-30] and women [31-33],psychosocial assessments are routinely conducted by fewerthan 20% of prenatal care providers [34]. In systems with-out linkages between assessment, referral, and mentalhealthcare, only 18% of pregnant and postpartum womenwho are assessed as having mental health problems actuallyfollow up with a referral that they have been given [35], andfewer than 15% of those needing care receive some form oftreatment [35,36]. The problem is further complicated byevidence that most women do not voluntarily disclose men-tal health concerns [22,37,38] (despite the fact that <4%refuse provider-initiated assessment) [39,40]. The cycle ofunder detection and under treatment is perpetuated bya ‘catch 22’ where providers do not assess women be-cause no follow-up services exist [39], and because womenare not assessed, they are not referred and treated. Targe-ting the individual components of assessment, referral,or treatment in isolation will not address the need in thatit is not feasible to enhance psychosocial assessment with-out simultaneously increasing service capacity to receive

http://clinicaltrials.gov/ct2/show/NCT01901796


referrals. Improvements in psychosocial care can only beaddressed as an integrated process of assessment-referral-treatment.

Integrated perinatal mental healthcareIntegrated perinatal mental healthcare - the systematic link-age of assessment, referral, and treatment [41] - has beenrecommended by national bodies [25]. Integrated care is amore efficient approach to primary care management ofdepression and anxiety in that it improves access, adher-ence, and treatment response while being cost-effective[41-44]. Very few studies have evaluated integrated psy-chosocial care during the perinatal period [40,45,46]. Inthese studies, the high prenatal ‘screening’ rates of 95%[45] and 62.5% [40] and low refusal rates (<4%) demon-strate women’s acceptance of routine screening and follow-up care [40]. The predominant limitations of existing studiesof integrated perinatal mental care (and areas we aim toimprove upon) are: 1) all lacked a comparison group; 2) allprimarily targeted depression without addressing stressand anxiety; 3) most conducted a minimal feasibility as-sessment, providing little guidance for improving the inter-vention or understanding its most effective components; 4)none evaluated clinical outcomes; 5) none used techno-logical (for example, web-based) approaches to support in-tegrated care, although recommended as a key element ofsuccess of integrated care [41]; and 6) none targeted themost prominent barriers to mental healthcare reportedby providers (for example, lack of time to screen, lack ofscreening tools and knowledge regarding their use, lackof referral mechanisms, unavailable and inaccessible non-pharmacological therapies) [28,47,48] or by pregnant/postpartum women (lack of time, preference for workingthrough their symptoms on their own, stigma associatedwith treatment, inability to find/access/afford nonphar-macologic therapy) [22,24,37]. Together, these limita-tions highlight the lack of utility that current researchoffers in terms of implementing integrated psychosocialcare in clinical settings. There is a need to design andrigorously evaluate integrated interventions that reducebarriers and promote access to mental healthcare bylinking standardized psychosocial assessment to effec-tive mental healthcare.

Standardized psychosocial assessmentPsychosocial assessment comprises the use of a standard-ized screening tool (for example, Edinburgh Postnatal De-pression Scale, EPDS) in addition to a holistic assessmentof psychosocial risk factors (for example, Antenatal RiskQuestionnaire, ANRQ-R) [1]. Standardized psychosocialassessment is feasible [31,35,49,50], improves detection[51,52] and facilitates triaging of women by symptom se-verity to ensure that women receive appropriate services[1]. However, serious resource limitations (for example,

lack of time and assessment tools) constrain many pri-mary care providers from routine assessment of mentalhealth problems. Computer-based psychosocial assess-ment conducted in primary care can address such limita-tions. Evidence exists that patients and providers find theuse of computer-based screening acceptable and feasiblefor inquiring about sensitive issues, including prenatal[53] and postnatal intimate partner violence [54] andmental health [55,56]. It is also well-suited for busy clinicalsettings in that it offers consistency, is resource-sparing,can be tailored to meet the needs of patients, can be usedwith audio/video for low literacy, easily provides a real-time summary for patients/providers [56,57], achievessimilar rates of disclosure to written- or interview-basedscreening, and is preferred by patients due to its perceivedanonymity [56,58,59]. However, a recent systematic reviewdemonstrated that, on its own, assessment is ineffective inpreventing or treating depression [60] and others haveshown that it does not improve linkage with healthcare inthe form of follow-up assessment or treatment [21,61].Thus, in order for mental healthcare to be effective, psy-chosocial assessment must be systematically linked totreatment.

Cognitive behavioral therapyCognitive behavioral therapy (CBT) is a highly effectivetreatment for depression and anxiety [62,63]. Since pre-natal mental health problems are characterized by theco-occurrence of anxiety and depression [3,16,17], CBT(including online CBT) is recommended in national guide-lines as an early intervention for improving maternal-childoutcomes [25]. Randomized controlled trials (RCTs) ofgroup-based CBT for new mothers [64-68] and pregnantwomen [69,70] demonstrate that group CBT is acceptableand efficacious in reducing risk and symptoms of postpar-tum depression [64-68].However, individual- and group-based CBT are fre-

quently inaccessible by pregnant women due to longwait times (groups often small; number of therapistslimited) and expense (that is, often not covered by healthinsurance) [71]. Barriers that are unique to childbearingfamilies (for example, care of other children) can also hin-der sustainability of women’s attendance at individual-and group-based CBT sessions [69,72]. Furthermore, preg-nant women with mild and moderate symptoms may notbe offered CBT due to resource constraints within thehealthcare system that restrict these limited services towomen with severe symptoms who present with the great-est need at the current time. Consequently, women withmild and moderate symptoms are underserved. Withouttreatment, there is evidence that 48% of pregnant womenwith anxiety and 71% of those with depression continueto experience symptoms throughout the postpartumperiod [2], with as many as one-third of new mothers


experiencing symptoms up to 4 years postpartum [73,74].As such, the delay in not treating pregnant women withmild or moderate symptom severity can lead to substantialpersonal, societal, and system costs if their symptoms be-come chronic or more severe over time [75]. Accessibleand available mental healthcare is a priority for this vul-nerable population.Few trials have evaluated CBT in pregnancy [66,69,

70,76,77]. Pilot testing of a prenatal workbook-based CBTplus telephone coaching by members of our research teamrevealed four key findings: 1) pregnant women found theprogram acceptable and helpful; 2) they wanted CBT earl-ier in pregnancy; 3) they wanted an online format; and 4)they recommended shorter modules [78]. The proposedtrial incorporates these pilot results by using six, 30-minute modules (versus the original three), delivering theintervention early in pregnancy (first and second trimes-ter), and adapting the CBT workbook for online use with-out the use of a telephone coach.Online CBT is resource-sparing, clinically and cost-

effective, acceptable [79-82], and accessible [79], and hasbeen recommended for treatment of anxiety and depres-sion in primary care [83]. A meta-analysis reported thatonline CBT produces moderate to large effects, is as ef-fective as face-to-face CBT, and has lower attrition rates(20%) than group-based CBT (40 to 50%) [7,84]. Al-though not tested in pregnant women, online CBT is anideal treatment because it can overcome major deter-rents to mental healthcare cited by pregnant/postpartumwomen, including: long wait times [35], inaccessibility[35], lack of time [35], finding childcare [24,85], stigmaof attending care [24], and treatment expense [35]. OnlineCBT satisfies the majority (93%) of distressed women’spreference for self-help [24] and should improve aspectsof psychological health (for example, mastery, resilience)related to poor pregnancy outcomes [86]. Importantly, on-line CBT can be embedded in current delivery systems,creating a sustainable approach to effective perinatal men-tal healthcare. Finally, evidence exists that online CBT andonline CBT plus telephone [87] or email [88] support by apsychologist are equally efficacious in reducing depressionand promoting adherence. Thus, online CBT, as a stand-alone intervention, offers a highly cost-effective approachto mental healthcare that is independent of limited humanresources.

Cost-effectivenessThe cost-effectiveness of integrated perinatal mentalhealthcare has not been evaluated [68,89]. However, aneconomic evaluation of the cost of treating postpartumdepression demonstrated that public health costs weretwice as high in women with postpartum depressioncompared to those without depression [75]. At a preva-lence rate of 25% among childbearing women, prenatal

mental health problems pose a substantial economic andhuman resource burden to the healthcare system. How-ever, widespread implementation of integrated prenatalmental healthcare (even resource-sparing approaches)will require a substantial commitment of resources, andan economic evaluation that considers the individual(maternal, family, child), local (clinic-and community-based), and societal implications of early, prenatal inter-vention compared to usual prenatal care is essential.

Mechanisms of integrated psychosocial careIntegrated prenatal psychosocial care is a complex inter-vention with several components. We found no studiesthat described mechanisms by which prenatal interven-tion led to improved outcomes [90]. As noted in theMedical Research Council Framework for Complex In-terventions, without this knowledge it is difficult to de-fine which components (for example, program content,intervention characteristics, method of delivery, assess-ment approach, referral processes) of an interventioncontribute to its impact and should be replicated inother settings. Given the need for widely accessible in-terventions across a diverse spectrum of perinatal careproviders and settings (midwives, nurses/nurse practi-tioners, family physicians, obstetricians), it is critical toidentify the key components of the integrated interven-tion that contribute to its effectiveness and facilitate suc-cessful implementation across settings. In practice, apregnant woman would complete the brief online psy-chosocial assessment while waiting for her clinic ap-pointment, and her perinatal provider would accessthese results in ‘real time’ online (for example, a sum-mary of psychosocial risk plus question responses). Adecision-making algorithm would provide guidance onthe most appropriate referral options for the provider todiscuss with the woman. Thus, a key aspect of this studyis to understand what aspects of the intervention en-hance or deter from its implementation success and in-tegration into routine clinical practice.

Maternal-child outcomesStrong evidence exists supporting a deleterious, endur-ing effect of poor prenatal mental health on adverse fetal[91] and child outcomes [11,18]. Two decades of longi-tudinal research have demonstrated a clear, independentassociation between maternal prenatal distress andneurodevelopmental outcomes in children [11,18,92]and adolescents [93]. Although well established in ani-mal research, early human studies provide evidence ofvarious biological pathways underlying the link be-tween prenatal distress and infant/child outcomes, in-cluding epigenetic mechanisms (that is, fetal DNAmethylation, placental gene expression [91,94]), im-paired neurogenesis [95], and dysregulation of the fetal


hypothalamic-pituitary-adrenal (HPA) axis [96,97].However, the interplay of influences in the prenataland postnatal environments and, in particular, the ex-tent of the moderating effect of postnatal interventionon fetal development that has been impacted by pre-natal depression, anxiety, or stress is largely unknown.Together, this evidence implies that early prenatalintervention should be explored as a means to inte-rrupting the risk of prenatal distress on infant and childwell-being.Very few studies have evaluated the impact of prenatal

CBT on infant outcomes [76,98], and none have deter-mined the influence of integrated perinatal mentalhealthcare on infant or child well-being, maternal earlycaregiving practices, or the maternal-child relationship.Furthermore, a recent Cochrane review recommended aRCT to explore the value of integrated prenatal psycho-social care on maternal-child outcomes [99]. This is animportant line of inquiry, given that symptoms of prenataldepression, stress, and anxiety tend to continue into thepostnatal period, influencing the quality of the child’spostnatal environment [7,8,73]. From a healthcare systemand societal perspective, the costs associated with poorpregnancy outcomes are substantial [11,18,75,100-102],and treatment options for postpartum mental health [36]and child developmental problems [103] are severelylimited. Thus, there is a need to evaluate whether earlyprenatal intervention can prevent or lessen the risk ofadverse maternal-child outcomes.

The need for a trialImprovement in perinatal mental healthcare mustbegin with an integrated, feasible, effective, and resource-sparing approach to routine prenatal psychosocial assess-ment that is seamlessly linked to referral and treatmentacross the perinatal period. A systematic review that isbeing conducted by our research team on the effective-ness of non-pharmacological prenatal interventions onmaternal-child outcomes identified four major gaps inprenatal mental health intervention research involvingthe lack of evaluation of: 1) the clinical effectiveness ofintegrated psychosocial assessment, referral, and care;2) the impact of prenatal interventions on prenatal dis-tress or infant/child outcomes (that is, the main out-come has primarily been postpartum depression); 3)factors that contribute to the effect of prenatal interven-tions (that is, how they worked); and 4) the cost-effectiveness of prenatal interventions. The current studyaims to address these four gaps by determining theclinical- and cost-effectiveness of integrated psycho-social assessment-care-referral compared to usual pre-natal care, evaluating process outcomes of integratedpsychosocial care, and describing the determinants ofeffectiveness of integrated psychosocial care.

Research questionsThe research objectives, primary and secondary researchquestions, and hypotheses associated with the four iden-tified knowledge gaps are described in Tables 1 and 2.

Methods/DesignDesignThe proposed study is a randomized, controlled superior-ity trial of two parallel groups that includes a prospectiveeconomic evaluation (Figure 1). It has two phases, inclu-ding: Phase 1 - a randomized controlled trial designed toevaluate the clinical and cost-effectiveness of an integratedpsychosocial assessment-referral-CBT intervention; andPhase 2 - a qualitative descriptive component designed toassess the efficiency, utility, usability, feasibility, acceptabil-ity, and mechanisms of the intervention (that is, the activeingredients within the intervention and how they exerttheir effect) [104]. The research design best suited for an-swering questions regarding effectiveness, mechanisms,and acceptability/feasibility is a design combining a RCTwith a qualitative component [105].

Randomized controlled trialSetting and recruitment proceduresRecruitment will take place at four primary care clinicsin two large, urban Canadian cities. Two of the clinicsprimarily serve an ethnically diverse, socioeconomicallydisadvantaged population, with the remaining two clinicsserving a largely middle class, Caucasian population.Family physicians in these clinics provide complete andshared prenatal care. Women under complete care re-ceive all prenatal and delivery care from the family phys-ician, while those under shared care receive care from afamily physician up to 28 weeks gestation and from anobstetrician thereafter. Family physicians providing careat these clinics do not have specialized mental healthtraining. Eligible women arriving for their prenatal careappointment will be invited to participate in the studyby clinic administrative staff. Clinic staff will give womenwho are interested in study participation a tablet with alink on the main screen to the consent and question-naire. A research assistant will be available to answerquestions about the study. Women agreeing to studyparticipation will complete the electronic consent on acomputer tablet.

Participant eligibilityPregnant women are eligible to participate if they are:1) <28 weeks gestation (Note: The upper limit of <28 weeksallows time to complete six modules and follow-up ques-tionnaires prior to delivery); 2) able to speak/read English;3) willing to complete email questionnaires; 4) have no,low, or moderate psychosocial risk on screening with theDepression, Anxiety, and Stress Scale at recruitment

Table 1 Primary objective, research question, and hypotheses

Gap objective Research question Testable hypotheses

1 To compare the clinical effectiveness ofintegrated psychosocial assessment-care-referral versus usual prenatal care on prenataldepression, anxiety, and stress symptoms

What is the effect of integrated, onlinepsychosocial care delivered in pregnancy towomen with low or moderate psychosocialrisk on the presence and severity of prenataldepression, anxiety, and stress symptoms at 6to 8 weeks post-randomization compared tousual prenatal care?

Presence of symptoms: Compared to womenin the control group, fewer women in theintervention group will have depression, anxiety,and stress symptoms (for example, be above theestablished cut-off for the DASS21 and EPDS).

Severity of symptoms: Women in theintervention group will have lower severity ofdepression, anxiety, and stress (that is, they willhave lower mean scores on the depression,anxiety, and stress subscales) compared to thosein the control group.


(Table 3); and 5) are eligible for CBT (Table 3). As inmany trials of CBT [68], women on antidepressantswill not be excluded because the study objective is notto compare CBT and pharmacological therapy (we plana subgroup analysis - see Analysis). We are includingwomen with ‘no’ symptoms of psychological distress atthe time of recruitment because up to 25% of womenin this group may develop mild or moderate symptomsof depression, anxiety, or stress during the course ofthe trial [106]. In addition, it is important to follow-upsubjects who were not identified as having depression,anxiety, or stress symptoms on recruitment in order tounderstand the implications of false positives (that is,women identified by screening as positive, but who donot have symptoms of depression, anxiety, or stress)and false negatives (that is, women not identified byscreening, but who do have symptoms of depression,anxiety, or stress) on the cost-effectiveness of the inter-vention package as a whole.

PrerandomizationAll women will complete the Depression, Anxiety, and StressScale (DASS21) and the Edinburgh Postnatal DepressionScale (EPDS) on recruitment (prerandomization) to deter-mine study eligibility and provide baseline data on theirmental health status. A computer algorithm designed forthis study (Table 3) will calculate symptom scores of theDASS21 and EPDS. If a woman’s scores indicate that sheis high risk, a computer-generated message will thank herfor her study participation and indicate that the researchnurse will contact her. Then, the software program willgenerate an email to the research nurse who will access thewoman’s online assessment results, telephone the womanto inform her that she is excluded from the study, providefeedback on her assessment and, with permission, arrangeappropriate referrals.

PostrandomizationWomen in the intervention group will complete theAntenatal Risk Questionnaire (ANRQ-R). Those who are‘unsuitable for CBT’ (Table 3) based on the established

algorithm using the ANRQ-R and the EPDS will also be ex-cluded from the trial. They will receive a computer messagethanking them for their participation and informing themthat the research nurse will be contacting them. Within 24hours, the research nurse will contact these women,inform them of their ineligibility, provide feedback on theANRQ-R, and, with permission, set up a referral with theirhealthcare provider. The research nurse will document allreferrals made in a computer-based tracking system devel-oped and tested through previous screening trials.As part of the safety protocol, women in the control and

intervention groups assessed as having high symptom scoresat any follow-up assessment based on the DASS21 andEPDS will be referred by the research nurse to their health-care provider (Table 3). All participants assessed as high riskat a follow-up assessment will remain in the trial for the pur-pose of answering follow-up questionnaires. This is criticalto determine whether the integrated intervention facilitateslinkage to mental healthcare. Given the stability of untreatedmental health symptoms across time [2], we anticipate thatan extremely small proportion (4 to 10%) of participants willbecome high risk after initially being assessed as ‘no’ or ‘lowto moderate’ risk at recruitment [14,31].

Randomization and allocation proceduresOnce eligible women complete the consent to partici-pate on the tablet, a simple computer random numbergeneration algorithm (1:1 allocation ratio) designed forthe study by the Clinical Research Informatics Core atthe Women’s and Children’s Health Research Institute(University of Alberta) will automatically randomizewomen. This will be followed immediately by a computer-generated message notifying women of their group assign-ment. The computer randomization ensures that theresearch assistant at the clinic is unaware of the participant’sgroup assignment prior to allocation, and thus allocationconcealment is maintained.

Sample size estimationThe sample size calculation is based on the primary out-come of symptoms of depression, anxiety, and stress as

Table 2 Secondary objectives, research questions, and hypothesesGap objective Research question Testable hypotheses

1-2 To compare the clinical effectiveness of integratedpsychosocial assessment-care-referral versus usualprenatal care on postnatal mental health,psychosocial resources, infant health, and family health

Compared to usual care, what isthe effect of integrated, onlinepsychosocial care delivered inpregnancy on:

Compared to women in the control group,those in the intervention will have significantly:

-decreased presence and severity of depression,anxiety, and stress symptoms at 12 weekspostpartum;

…..prenatal and postpartummental health?

-increased psychosocial resources (self-efficacy,mastery, self-esteem, coping); improved sleepquality; and higher relationship quality at 6 to 8weeks postrandomization and 3, 6, and 12months postpartum.

…infant health? Infants of women in the intervention group willhave significantly higher: 1) 5-minute Apgar scores,2) birth weight, 3) gestational age, 4) maternal-child attachment, and 5) significantly reduced‘dysfunctional’ infant behavior compared to theintervention group.

…family health? The intervention group will have significantlyhigher parenting competence and partnerrelationship quality and significantly lowerparenting stress compared to the control group.

To evaluate process outcomesof integrated psychosocial care

Is integrated psychosocial caremore efficient, feasible, andacceptable than usual prenatal care?

Efficiency: Compared to the control group, asignificantly higher percentage of women in theintervention group will have a psychosocialassessment and receive treatment. Theintervention group will have significantly lowerpercentage of women receiving emergencymental healthcare compared to the control group.

Feasibility: ≥ 90% of providers and women reportpsychosocial assessment is easily done as a componentof routine prenatal care, ≥ 95% of intervention groupwomen will access cognitive behavior therapy modules(CBT) within 2 weeks of psychosocial assessment, ≥ 80%of intervention group will access the CBT modules every1-2 weeks, ≥75% intervention group will complete allCBT exercises, and intervention group will complete 80-100% modules within 6 to 8 weeks.

Acceptability: ≥ 90% of intervention group womenand providers will report tablet-based psychosocial as-sessment during prenatal care acceptable, > 90%women will report that they could provide ‘honest’ re-sponses, and ≥ 90% of intervention group women andproviders will find the CBT modules acceptable.

Utility: ≥ 85% of intervention group will report thatthe CBT homework exercises were useful, and ≥ 90%of intervention group will report each module asuseful.

Usability: ≥ 90% of intervention group will report thatthe exercises and modules were clear, easy tounderstand, and easy to navigate around.

(Note. Targets are based on meta-analyses ofadherence and satisfaction rates [44]).

3 To describe mechanisms of integrated care What are the mediators andmoderators of the interventioneffect?

Psychosocial resources (self-efficacy, mastery, self-esteem, coping), sleep, and relationship quality willmediate the impact of the intervention on maternal,child, and family outcomes; and participantcharacteristics will moderate the effect (for example,demographics, use of antidepressants).

4 To compare the cost-effectiveness of integratedpsychosocial care compared to usual care

Is the integrated psychosocial caremodel cost-effective whencompared to usual prenatal care?

The expected incremental cost effectiveness ofintegrated psychosocial assessment, referral, andtargeted cognitive behavioral therapy is costeffective at values of health considered acceptablein the Canadian healthcare system.


Figure 1 CONSORT Trial Flow Diagram. DASS21, Depression, Anxiety, Stress Scale; EPDS, Edinburgh Postnatal Depression Scale.


measured by the Depression, Anxiety, and Stress sub-scales of the DASS21 [107]. We calculated the samplesize required to test the minimum clinically importantdifference in each subscale and selected the highest onefor the final sample size. Based on DASS21 data col-lected as part of Australia’s national perinatal mentalhealth initiative, standard deviations for the depression,anxiety, and stress subscales in pregnant women are 5.4,10.2, and 8.6 [108]. To determine the minimal clinicallyimportant difference, we used Milgrom et al.’s [66] ap-proach for calculating the difference in scores on eachsubscale that would shift a woman one level of severity -the minimal, reasonable expectation for an effective the-rapy. For example, the DASS21 manual ‘categorizes’women as having normal, mild, moderate, severe, and ex-tremely severe symptoms of depression, anxiety, and stress[107]. To shift women from ‘mid-range’ moderate to mildseverity on the depression, anxiety, and stress subscales

would require a reduction of 4 points in each subscale.Therefore, based on the sample size formula for compari-son of two means (2-tailed) at a significance level of 5%(1.96), a power of 80% (.84), and a minimally clinically im-portant difference of 4 points, 204 women with mild tomoderate symptoms of psychological distress are requiredin the trial (Table 4).Based on a 25% prevalence rate of low-moderate pre-

natal psychological distress [1,31], a final sample size of816 eligible women (408 per group) would be needed.This corresponds to a moderate effect size (d = 0.4 to 0.7)across subscales, which is consistent with a meta-analysisof effect sizes of online CBT [84]. This sample size isalso adequate to conduct structural equation modelingto address secondary objectives related to mechanismsunderlying the impact of the intervention. Kline rec-ommends a minimum sample size of 200 for complexstructural equation models or 10 to 20 cases per level

Table 3 Criteria for ‘high risk’ and referral to physician

Based on baseline DASS21/EPDS Based on ‘unsuitability’ for CBT (intervention group)

Women with ‘severe’ or ‘extremely severe’ psychological distressbased on one or more of the following criteria:

Women in intervention group with three or more of the following criteria:

1. Depression subscale ≥21 and/or 1. ANRQ-R positive for childhood emotional neglect, childhood emotionalabuse, or childhood sexual or physical abuse and/or

2. Anxiety subscale ≥15 and/or 2. ANRQ-R positive for multiple major stressors (for example, major financial issues, bereavement, or separation)

3. Stress subscale ≥26 3. Current substance use or domestic violence

4. EPDS positive Q10 (1, 2, or 3) 4. EPDS positive Q10 or total EPDS score >15

ANRQ-R, Antenatal Risk Questionnaire-Revised; CBT, cognitive behavior therapy; DASS21, Depression, Anxiety, Stress Scale 21; EPDS, Edinburgh PostnatalDepression Scale.


of a variable [109]. Therefore, this final sample size of816 participants is adequate to address primary andsecondary research questions. Accounting for a partici-pation rate of 50% based on previous studies of CBTin pregnant women [110], the exclusion of 15%of women who do not meet study criteria (5% highpsychosocial risk [1,31], 10% non-English speakingwomen), a conservative attrition rate of 35% based onprevious studies of prenatal CBT [84], and a 5% loss-to-follow-up (no data reported but our questionnairefollow-up through email should be largely unaffectedby change in residence etcetera), 1,673 women wouldneed to be invited to participate in the study to achievethe final sample size. Given that the number of newpregnant patients across the four recruitment sites is120 per month, the duration of recruitment is antici-pated to be 14 months.

InterventionThe intervention consists of usual prenatal care plus anintegrated intervention comprising: 1) online psycho-social assessment, 2) referral, and 3) online CBT.

Table 4 Sample size estimation

N = 2(0.84 + 1.96)2 * (σ/§)2

σ = standard deviation of the primary outcome (Depression, Anxiety,Stress subscales of DASS21)

§ =minimal clinically important difference

Depression subscale Anxiety subscale Stress subscale

N = 2(0.84 + 1.96)2 *(σ/§)2

N = 2(0.84 + 1.96)2 *(σ/§)2

N = 2(0.84 + 1.96)2 *(σ/§)2

N = 2(0.84 + 1.96)2 *(5.4/4)2

N = 2(0.84 + 1.96)2 *(10.2/4)2

N = 2(0.84 + 1.96)2 *(8.6/4)2

N = 28.6 per group N = 102 per groupa N = 72.5 per groupaLargest sample size per group = 102. Total sample of women with mild tomoderate psychological distress = 204. Based on a prevalence rate of 25% oflow-moderate symptoms in pregnant women, a final sample size of 816(408 per group) will be required (204/N = 25/100).

Online psychosocial assessmentFollowing randomization to the intervention group, par-ticipants will begin the intervention by self-completingthe psychosocial assessment (ANRQ-R) on the tabletwhile waiting for their prenatal appointment. Developedand psychometrically tested by one of our team mem-bers [31] the ANRQ-R was designed to be embeddedwithin an integrated system of assessment-referral-careto identify psychosocial risk factors associated with poormental health outcomes in pregnant women. A systemfor categorizing level of risk and tailored referral hasbeen devised to enable triaging of women to appropriateservices (including CBT) [1]. Its recommended usealongside the EPDS permits identification of psycho-social risk and current symptoms (past 7 days). Both in-struments can be completed together in less than 10minutes. The ANRQ-R has high levels of acceptabilityand satisfactory psychometric properties (sensitivity 0.62;specificity 0.64) [1,31], comparable to other commonlyused self-report depression/anxiety tools. The EPDS is awidely used 10-item self-report depression scale used todetect depression symptoms during the previous 7 days[111]. Psychometrically validated for use in pregnant andpostpartum women [112], testing revealed sound psy-chometric properties (sensitivity 86.7%; specificity 78%;positive predictive value 74%, α = .87) [111]. An intro-ductory section to the ANRQ-R describes the import-ance of routine psychosocial assessment and how it willbe used as an initial step to help women with emotionalhealth concerns.

ReferralOnce women in the intervention group submit theirpsychosocial assessment, a software program developedfor this study will use a scoring algorithm to determinewhether the intervention group participant meetscriteria for CBT based on ANRQ-R and EPDS scores(see Participant Eligibility). A trained research nurse will re-ceive notification of participant enrolment and telephoneall women in the intervention group within 48 hoursto review the results of the psychosocial assessment. A


standardized approach will be utilized to review the re-sults; however, this review will be tailored to participants’specific needs in response to their questions and concerns.The research nurse will refer women who meet criteria tothe CBT modules (for example, give them the password,web-link, and instructions on accessing and completingthe modules).Women in the intervention group who were assessed as

‘no’ risk on recruitment, but who convert to low or mod-erate risk at the 6 to 8 week postrandomization follow-upwill also be referred to the CBT modules. In this case, theresearch nurse will receive an automatically generatedemail and will follow-up with a discussion on symptomscores and referral to the CBT modules. Including womenin the study with a gestational age of less than 28 weeksgives sufficient time for women detected at 6 to 8 weekspostrandomization to complete the intervention duringpregnancy.

Online cognitive behavior therapyEligible women in the intervention group will be asked tocomplete the six, 30-minute online, interactive CBT mod-ules over 6 to 8 weeks [113,114]. Four [77] to six [69]CBT sessions have been found to effectively reduce de-pression symptoms. The topics of the modules are: 1) tak-ing stock; 2) identifying and labeling emotional healthconcerns; 3) changing distorted thinking; 4) understandingand changing actions, responses, and behavior; 5) relax-ation; and 6) developing and maintaining a plan. Eachmodule has interactive homework assignments thatwomen complete online. Each assignment has one to fouroptions and women select the one (or more) that bestsuits their needs. Completion of the homework is requiredbefore progression in the modules can occur. The mod-ules utilize pregnancy-relevant scenarios and these areused as the basis of examples in the homework assign-ments. The online delivery allows women to set their ownpace by completing the modules at a time and locationthat is most convenient and ensures standardization of theintervention. Women will access the modules using ausername and password, and content that women providein the homework assignments is accessible only by them.

Comparator: usual prenatal careThe control group will receive usual prenatal care. Usualprenatal care at the study sites does not include routinepsychosocial assessment or follow-up of psychosocialconcerns. Given that this typifies standard prenatal carein the majority of perinatal settings in North America,‘usual prenatal care’ is the best comparator. All women inthe control group will be followed up at 6 to 8 weeks post-randomization and at 3, 6, and 12 months postpartumusing the same questionnaires as those delivered to theintervention group. An automatic email will be generated

to the research nurse for women in the control group whoconvert from no, low, or moderate risk to high risk at the6 to 8 week postrandomization follow-up. The researchnurse will contact these women and, with permission,set up a referral to their healthcare provider. Thesewomen will continue in the trial to complete all follow-up questionnaires.

Definition and measurement of outcomesPrimary outcome The primary outcome is the presenceand severity of current prenatal depression, anxiety,and stress symptoms as measured by the DASS21 [107](Table 5). The DASS21 has been widely used and psycho-metrically tested; it distinguishes well between symptoms ofdepression, anxiety, and stress in clinical and non-clinicalpopulations [107,115,116]. It is used in clinical settings toscreen pregnant and postpartum women for presence andseverity of current symptoms of depression, anxiety, andstress [108,117]. The DASS21 has good psychometric prop-erties with Cronbach α’s of 0.91, 0.80, and 0.84 respectivelyfor depression, anxiety, and stress subscales [116]. Highcorrelations with other standardized depression, stress, andanxiety measures (for example, Beck Depression Inventory,State-Trait Anxiety) and clinical assessments demonstrateits validity [118,119].The presence of symptoms of prenatal depression, anxiety,

and stress is measured as the proportion of women scoringabove established cut-offs (>10; >8; >15, respectively) [107].Severity of symptoms is measured by the mean depression,anxiety, and stress scores. Ranges of scores correspondingto symptom severity levels of ‘no’, ‘mild’, ‘moderate’, and‘severe’ are also well-established through psychometrictesting: depression (none 0-9; mild 10-13; moderate 14-20;severe > 21); anxiety (none 0-7; mild 8-9; moderate 10-14;severe > 15); and stress (none 0-14; mild 15-18; moderate19-25; severe >26) [107].

Secondary outcomes All secondary outcomes and theirmeasures are described in Table 5. The secondary clin-ical outcomes are: presence and severity of symptoms ofpostpartum depression, anxiety, and stress [107]; prenataland postnatal self-efficacy [123], social support [121],sense of mastery [122], self-esteem [124], sleep [125,126],relationship quality [7,129], coping [130], and resilience[91]; 5-minute Apgar score; gestational age; birth weight;maternal-infant attachment [131]; infant behavior [132];infant development [91,133]; and parenting stress/compe-tence [127,128] (Table 5). These outcomes were selectedbecause of their association with maternal depression,anxiety, and stress and their potential modifiability by theintervention. Other secondary process outcomes relatedto the intervention include its: cost-effectiveness; effi-ciency; utility; usability; acceptability, and mediators andmoderators of effect (Table 5).

Table 5 Data collection schedule and measures

Variable (Measure) Timing of measures

Baseline 6 to 8 weeks post-randomization(pregnancy)

3 monthspostpartum

6 monthspostpartum

12 monthspostpartum

PHASE I

Demographics (education, income, maternal age at recruitment,ethnicity) (Items from Maternity Experiences Survey, bMES [120])

X

Obstetric and medical history (parity, chronic and pregnancycomplications, type of delivery, weight - pre-pregnancy, delivery, 6weeks postpartum) (Items from MES)

X X

Mental health history (history of depression, anxiety, stress; age ofonset of previous episodes of mental health problems) (Items fromMES)

X

Pharmacologic therapy for depression/anxiety (past; current) (Itemsfrom Canadian Community Health Survey, CCHS)

X X X X X

Social support (Interpersonal Support Evaluation List, ISEL [121]) X X X X X

Prenatal depression, anxiety, stress symptoms (Depression, Anxiety,and Stress Scale, DASS-21 [107] - presence (percent above cut-offpoint) and severity (mean score, standard deviation)

X X

Postnatal depression, anxiety, stresssymptoms (Depression, Anxiety, and Stress Scale, DASS-21 [107] -presence (percent above cut-off point) and severity (mean score,standard deviation)

X X X X X

aPsychosocial assessment (Antenatal Risk Questionnaire-Revised,ANRQ-R; includes substance use and violence) [1,31]

X X X X X

Depression (Edinburgh Postnatal Depression Scale, EPDS) [111] X X X X XaANRQ-R acceptability X

Mastery (Pearlin’s Mastery Scale) [122] X X X X X

Self-efficacy (Generalized Self-Efficacy Scale) [123] X X X X X

Self-esteem [124] X X X X X

Resilience (Connor-Davidson Resilience Scale) [91] X X X X X

Sleep (Pittsburgh Sleep Quality Index) [125,126] X X X X X

Parenting competence (Parenting Sense ofCompetence Scale, PSCS; subscales Efficacy, Interest, Satisfaction) [127]

X X X

Parenting stress (Parental Stress Scale) [128] X X X

Relationship quality and adjustment (Dyadic Adjustment Scale, DAS-7)[7,129]

X X X X X

Coping (Brief Cope) [130] X X X X X

Maternal-infant attachment (Condon and Corkindale) [131] X X X

Infant behavior (Infant Behavior Questionnaire) [132] X X X

Infant development (Ages and Stages Questionnaire,3rd edition, ASQ-3; The Baby Pediatric SymptomChecklist for Social/Emotional Screening) [91,133]

X X X

Birth weight (medical record) X

Gestational age (medical record) X

5-minute Apgar score (medical record) X

Other factors related to infant outcomes: feeding method (medicalrecord and parent-report); neonatal/infant health (medical record andparent-report) (Parent report items from the All Our Babies birthcohort studyc)

X X X

Patient diaries [134] (For economic analysis - including health serviceuse, medication use, productivity loss, personal cost)

X X X X X


Table 5 Data collection schedule and measures (Continued)

Quality of life (For economic analysis - SF-36,SF-6D to calculate QALY)[135]

X X X X X

Efficiency of intervention (percent of women with psychosocialassessment, referral, and care in IG versus CG; self-report and medicalrecord)

X X X

Utility of intervention (one question asked at the end of eachcognitive behavior therapy (CBT) homework exercise: This exercise wasuseful to me with four response options of I strongly agree, Isomewhat agree, I somewhat disagree, I strongly agree; one questionasked at the end of each CBT module: The information in this modulewas useful to me with same response options)

X

Usability of intervention (one question asked at the end of each CBThomework exercise: This exercise was clear and easy to understand withresponse options; 2 questions asked at the end of each module: 1) Theinformation in this module was clear and easy to understand; 2) It was easyto work through the module (for example, it was easy for me to get from onepart to the other, easy to find what I needed) with same response options)

X

Acceptability: Tablet-based psychosocial assessment (one question atend of completing ANRQ-R: I would recommend a tablet-based ap-proach to asking about emotional health to a pregnant friend with fourresponse options of I strongly agree, I somewhat agree, I somewhatdisagree, I strongly agree)

X

Acceptability: CBT (one question at end of each CBT module: I wouldrecommend this module to a pregnant friend who was struggling withstress, depression, or anxiety with 4 response options of I stronglyagree, I somewhat agree, I somewhat disagree, I strongly agree)

X

Overall assessment (two open-ended questions at the end of everyCBT module: 1) The thing I liked most about this module was….; 2) Thething I liked least about this module was….)

X

Log of interactions with participants (completed by research nurse) X X X

PHASE 2

Efficiency (Providers’ views of the efficiency of the process of clinic-based online psychosocial assessment)

X

Utility (Women’s views of how useful the modules in were in meetingtheir needs)

Usability (Women’s views of how easy/difficult the modules were tonavigate)

X

Feasibility (providers’ views of feasibility of conducting integratedintervention in their setting; women’s views of the feasibility of doingthe modules; Google Analytics for example, percent of womenaccessing CBT within 2 weeks postassessment; percent of womenaccessing each CBT module within 1 to 2 weeks; percent completionof all six CBT modules; percent completion of CBT modules within 8weeks)

X

Acceptability (women’s and providers’ views of acceptability/ability topromote disclosure)

X

Mechanisms (women’s views of why and how the intervention did/did not improve outcomes; how the intervention benefitted/did notbenefit them)

X

aIntervention group.bThe Maternity Experiences Survey (MES) is a national survey designed and administered by the Public Health Agency of Canada and Statistics Canada. The surveywas designed through an exhaustive process involving discussion, consultation, literature reviews, focus group testing, and two pilot studies [120].cThe ‘All Our Babies Birth Cohort’ study is a pregnancy birth cohort in Alberta, Canada. Details of the study methodology and design have been previouslypublished [136].


Data collectionProcedures The five data collection points for all studyparticipants are: recruitment; 6 to 8 weeks postrandomi-zation; and 3, 6, and 12 months postpartum (Table 6).On recruitment, all consent and baseline data are

collected via a computer tablet while women wait for theirprenatal appointment. Follow-up questionnaires will becompleted online. Participants will receive an email with apassword and link to the questionnaire on the projectwebsite (www.yourhope.ca). Retention will be enhanced

http://www.yourhope.ca

Table 6 Schedule of enrollment, interventions, and assessments

Enrollment Allocation CBT Suitability Post-randomization

TIME POINT -t1 0 0 T1(Baseline)

T2 (6 to 8 weeks

postrandomization)

T3 (3 months

postpartum)

T4 (6 months

postpartum)

T5 (12 months

postpartum)

ENROLLMENT:

Eligibility screen (based onDASS21 and EPDS)

X

Informed consent X

Allocation X

Determination of suitability forCBT (based on ANRQ-R)

X

INTERVENTIONS:

Psychosocial assessment(ANRQ-R)

X

Referral X

Online cognitive behaviortherapy

X X

ASSESSMENTS:

Baseline variablesa X

Primary outcome: Depression,anxiety, stress symptoms

X X X

Secondary outcomes -maternalb X X

Secondary outcomes -maternaland infantc

X X X

Utility, usability, acceptability ofintervention

X

Phase 2: Qualitative interviews X X X

ANRQ-R, Antenatal Risk Questionnaire; CBT, cognitive behavior therapy; DASS21, Depression, Anxiety, Stress Scale; EPDS, Edinburgh Postnatal Depression Scale.aBaseline variables: demographics; history-obstetric, medical, mental health diagnosis and treatment; social support; mastery; self-efficacy; resilience; sleep; partnerrelationship; coping.bSecondary outcomes - maternal: mental health treatment; social support; mastery; self-efficacy; resilience; sleep; partner relationship; coping; mental healthservice utilization.cSecondary outcomes - maternal and infant: mental health treatment; social support; mastery; self-efficacy; resilience; sleep; partner relationship; coping; parentingcompetence; parenting stress; maternal-infant attachment; infant behavior; infant development; gestational age; birth weight; 5-minute Apgar Score; mental healthservice utilization; quality of life.


using Dillman’s approach [137], where women who havenot completed the questionnaires within 1 week will re-ceive computer-generated email/smartphone remindersat 1, 3, 7, 10, and 14 weeks by Checkbox Survey Server.We will track reasons for nonadherence (for example,lost to follow-up).

Management No data are stored on the tablets; rather,when women ‘submit’ their information it is sent to asecure server housed in the Faculty of Medicine &Dentistry’s Data Centre (University of Alberta). Datatransfer between the tablet and server will be encrypted.Follow-up questionnaires will be distributed and submit-ted via email that is also encrypted. All processes involvingelectronic data capture and storage are managed by theWomen’s and Children’s Health Research Institute In-formatics Core at the University of Alberta. Once data col-lection has been completed, de-identified data will bestored in the Health Research Data Repository at the

University of Alberta. Access to the Repository is re-stricted to research team members conducting analyses.

Attrition, adherence, fidelity, and concomitant careAttrition rates in online CBT are roughly half those [84]of group-based CBT [68]. We will compare attritionrates in the intervention and control groups and conducttelephone interviews with women who drop out of theintervention group to assess reasons. Adherence, that is,the extent to which women complete the psychosocialassessment and CBT components, will be documentedthrough Google Analytics and application specific ana-lytics developed for this study (for example, numbermodules completed, length of time to complete modules,etcetera). As part of the qualitative descriptive compo-nent we will seek women’s opinions about aspects of thepsychosocial assessment that were challenging and fea-tures of the CBT modules that impacted their ability,need, or desire to complete them.


To improve adherence of women in the interventiongroup, the research nurse will outline the importance ofregular progress through the module homework and thebenefit of completing all modules when she provides in-structions on how to access the modules. A second keystrategy to optimize adherence is the use of email andsmartphone reminders to complete the CBT modules ifwomen have been inactive on the site for more than 2weeks. These reminders will automatically be generatedby the software program.The online format of the intervention preserves its fidel-

ity (that is, consistency in its components and delivery) andthus enhances external validity. To limit co-interventionbias, all women will be discouraged from participating inother self-referred forms of non-pharmacological mentalhealthcare. However, if the blinded physician detects symp-toms of psychological distress in the course of usualprenatal care, the study participant may initiate the rec-ommended pharmacological or non-pharmacologicaltherapy. Follow-up questionnaires will ask women todisclose any pharmacological or non-pharmacologicaltherapy that they have begun and this additional inter-vention will be accounted for in the analyses.

Minimizing the risk of biasBlinding We will ask women not to share their studyinvolvement with their physician in order to maintainphysician blinding and limit the possibility that the phys-ician would change his/her approach to ‘usual prenatalcare’. If physicians show greater vigilance in their routineprenatal care as a result of clinic trial involvement, weanticipate that both control and intervention groupswould be affected. It is not possible to blind participantsdue to the nature of the intervention. Women will self-report on all outcomes with the exception of birth weight,gestational age, neonatal health and feeding method atbirth, and Apgar scores. These data will be abstractedfrom the medical record by a research assistant blinded togroup allocation, thereby limiting ascertainment bias.Given that current mental health may impact women’sperceptions of infant development and maternal-infant at-tachment, we will control for current mental health whenanalyzing these outcomes (for example, DASS21 assessessymptoms of depression, anxiety, stress past 7 days). Re-searchers will be blinded throughout the trial. Finally, thedatabase used for analysis will not include women’s alloca-tion assignment, and therefore the researcher and assist-ant conducting data analyses will be unaware of groupassignment.

Selection bias Selection bias will be limited by consecu-tive recruitment of women that allows every eligiblewoman to have an opportunity to participate in the study;however, selection bias will still be a potential factor

affecting external validity given the non-random selectionof our study sites and the exclusion of non-English speak-ing women. We also aim to limit threats to internal valid-ity by reducing attrition through the design of the easilyaccessible, online CBT program rather than a group pro-gram. Information bias is minimized due to the use ofstandardized tools and prospective data collection. Co-intervention may occur if women in the interventiongroup seek additional formal or informal help for de-pression/anxiety symptoms in addition to the interven-tion. We will measure additional service use to parseout the independent effect of the CBT component (forexample, subgroup analysis) and to inform the eco-nomic analysis. Contamination will be minimized by: 1)having all eligible women ‘do the same thing’ in the clinic(for example, use computer tablet); 2) delivering the inter-vention away from the physician office; 3) encouragingwomen not to discuss trial involvement with other pa-tients in the clinic; and 4) ensuring that only women allo-cated to the intervention group access the CBT modulesby having a password-protected entry.

Ethics considerationsThe study protocol was approved by the Human ResearchEthics Board at the University of Alberta. Following elec-tronic consent, all women and sites will receive an emailedcopy of the Participant Information Letter and Consent.

Safety protocolSeveral strategies have been implemented to monitorboth intervention and control group women’s psycho-social risk and ensure their safety throughout the trial.

Intervention group Mental health crisis contact infor-mation is visible on a sidebar of the online CBT modulesand remains accessible at all times. The sidebar also con-tains a statement encouraging women who feel worsethan when they started the intervention to contact theresearch team’s mental health nurse through a dedicatedemail link. This message indicates that research nursewill contact the woman within 24 hours. An algorithmwill guide the research nurse’s decisions regarding thelevel of help or referral (for example, to a mental healthexpert on the research team or woman’s physician) that isprovided. Women who were assessed as ‘no’ risk but con-vert to low or moderate risk at the 6- to 8-week postrando-mization follow-up will be contacted by the research nurseand referred to the CBT modules (see Referral). All interac-tions and decisions will be documented in a computer-based tracking system by the research nurse. The mentalhealth therapists and psychiatrists on the team are availablefor consultation.At the end of each CBT module, women will complete

question 10 of the EPDS (self-harm thoughts over past


week) as a ‘required’ response. An affirmative response(Q10 = 1, 2, or 3) will generate an automatic ‘pop-up’message with crisis contact information for the woman’simmediate use and an email to the research nurse. The re-search nurse will contact the woman within 24 hours toassess further and ensure that the woman is receiving helpfrom a healthcare provider. If not, the research nurse willlink her to additional resources as guided by an algorithmdevised for this study. A 4% affirmative response rate toQ10 of the EPDS has been reported [14]. These womenwill continue in the trial. A log of interactions and deci-sions will be maintained by the research nurse.

Intervention and control group For women who con-vert from no, low or moderate risk (on recruitment) to‘high’ risk (based on DASS21 and EPDS scores) at anyfollow-up point, an automatic email will be generated tothe research nurse. The research nurse will contact thewoman, describe her assessment results, and create a re-ferral to her healthcare provider with her permission.These women will remain in the trial to complete thefollow-up questionnaires. All interactions and decisionswill be documented in a log by the research nurse.

Training The mental health research nurse will attend a4-hour training session conducted by mental health expertson the research team regarding the use of the algorithm toguide decision-making and referrals, the availability of localmental health services, and techniques for assisting womenin crisis, including domestic violence. The research nursewill collaborate with the perinatal provider and relevantlocal agencies to provide support.

AnalysesEffectiveness of intervention We will use descriptivedata (frequencies and 95% confidence intervals, CI; meansand standard deviation, SD) to describe the sample. Wewill test for differences in baseline characteristics using t-tests (means) and chi-squared tests (%). We will assess dif-ferences proportions and mean scores of primary and sec-ondary outcomes at each follow-up point using chi-squareand t-tests, respectively. We will use an intention-to-treatanalysis. We will also use multivariable logistic regressionto determine predictors of outcomes and report relativerisks and 95% CIs. Multivariable regression models will bebuilt using variables that are associated with outcomes atP <0.10 on unadjusted analyses. Primary analyses will usea type I error of 5% as a criterion for statistical signifi-cance, while a more stringent alpha of 0.01 will be usedfor secondary outcomes to account for multiple testing.Because women will be starting the intervention at dif-ferent points in pregnancy we will control for gestation.We plan to conduct an exploratory analysis using strati-fied analyses to explore a priori subgroup differences of

intervention effect by: (a) number CBT sessions, (b)antidepressant use, (c) symptom clusters, (d) severity ofsymptoms of psychological distress (DASS21), (e) add-itional mental health service use, (f) participant character-istics, (g) mental health history, and (h) gestation at timeof recruitment. We expect the volume of missing ques-tionnaire data to be low due to the design of the electronicdata capture that requires data fields to be populated priorto progressing to subsequent questions. As such, we donot plan to conduct imputation of missing data.

Efficiency, utility, usability, and acceptability ofintervention In addition to analyzing the qualitative datathat are gathered through Phase 2 interviews, we will usedescriptive statistics (frequencies, proportions, means,standard deviations) to describe the efficiency of the inter-vention (for example, percent of women with psychosocialassessment, referral, and care in intervention group versuscontrol group) and intervention group women’s percep-tions of the utility (for example, rated usefulness of exer-cises and module), usability (for example, rated ease of useof exercises and module), and acceptability (rated willing-ness to recommend intervention) of the intervention(Table 3). We will also identify the main predictors ofthese intervention features through multivariable logisticregression and will include variables such as demographiccharacteristics (including parity), comfort with computertechnology, current DASS21 scores, history of mentalhealth problems, and use of ancillary mental health ser-vices. All independent variables that are related to eachfeature at P <0.10 will meet criteria for entry to the multi-variable models. Adjusted odds ratios and 95% confidenceintervals will be reported.

Mechanisms of effect of intervention We will deter-mine factors that influence acceptability and uptake ofthe intervention using multivariable logistic regression,as adjusted relative risks and 95% CIs. As a preliminarystep to model building, we will conduct unadjusted lo-gistic regressions with the criterion for entry into thefinal multivariable model being P <0.05. We will usestructural equation modeling (SEM) to describe the directand mediated effects of the intervention on outcomes.SEM is highly useful for describing complex pathways be-tween an intervention and outcomes that can inform howthe intervention has its effect. Consistent with rigorousSEM methodology [109], we will develop a priori models.We will analyze and refine the fit of the model based onrecommended model fit indices (for example, model chi-square, Bentler comparative fit index) and theoreticalplausibility of the pathways [109]. We will use maximumlikelihood estimation for estimation of means, variances,and covariances in order to retain records with missingdata in our analysis. We will also analyze qualitative data


regarding participants’ and providers’ perspectives on whyand how the intervention did/did not improve outcomes(Table 3).

Cost-effectiveness of intervention The economicevaluation will be a within-trial cost effectiveness ana-lysis comparing the integrated intervention ‘package’with usual prenatal care. The perspective of the primaryanalysis will be that of the health and social care budget.A secondary analysis will adopt a societal perspective in-corporating personal costs and productivity costs inaddition to the health and social costs associated with thedelivery of the intervention (for example, cost of equip-ment, salary of research nurse and clinic staff) and subse-quent service utilization by study participants. Directhealthcare utilization will be extracted from patient records.Patient Diaries will be completed at 3-month intervals fromrandomization to end-of-follow-up, to gather retrospectiveaccounts of other health and social care utilization, out-of-pocket expenses and productivity costs (Table 5). The SF-36 will be completed at the same time points as the PatientDiaries. The Patient Diaries will contain validated resourceutilization and productivity cost questionnaires such asPRODISQ [134]. The primary outcome measure for thecost effectiveness analysis will be the Quality Adjusted LifeYear (QALY). Utilities for the construction of QALYs willbe obtained from the SF-36 data using the SF-6D algo-rithm [135]. As the time horizon for the analysis is lessthan 12 months, discounting will not be required [138].We will report the incremental cost per QALY gained forthe integrated intervention compared to usual prenatalcare. Uncertainty in the expected costs and outcomes forthe integrated intervention and usual prenatal care will becharacterized using the non-parametric bootstrap. The re-sults of the bootstrap analysis will be used to constructscatterplots on the cost effectiveness plane and cost-effectiveness acceptability curves showing the probabilitythat the integrated intervention is a cost effective use ofhealthcare resources for a range of values of health.

Qualitative descriptive studyDesignIncorporating participant perspectives and experiences toassess the suitability and utility of interventions is critical intrials of complex interventions in order to identify compo-nents that may influence the outcomes [139]. Phase 2 uti-lizes a qualitative descriptive study to assess women’s andhealthcare providers’ views on efficiency, utility, usability,feasibility, acceptability, and potential mechanisms of actionof the intervention.

MethodsParticipant eligibility and recruitment All interven-tion group participants and healthcare providers working

at study sites are eligible for participation in Phase 2.Purposeful sampling will be used to maximize variabilityin the sample, ensuring that a broad range of views anddemographics are represented [140]. We plan to inter-view 15 to 20 intervention group women and 8 to 10providers (for example, nurses, family physicians) withthe final sample size determined by data saturation.Given the importance of understanding factors contrib-uting to attrition, we will also interview interventiongroup women who do not complete all CBT modules. Inorder to capture women who may not complete all sixmodules, a notification at the end of each of the fourth,fifth, and sixth CBT modules will invite women in theintervention group to participate in a follow-up inter-view. Selection of the affirmative response will generatean automatic email to the research coordinator forfollow-up. Emails distributed by each of the clinic man-agers will invite clinic staff members to participate in afollow-up interview.

Data collection and management We will conduct in-dividual face-to-face or telephone-based interviews.Semi-structured interview guides will be used [140] toask participants their views on the efficiency, utility, us-ability, feasibility, acceptability, and mechanisms of ac-tion of the intervention (Table 3), as well as itsstrengths, suggestions for improvement, componentsthat were effective/not effective, and the benefits thatthey experienced. Interviews are expected to take onehour and will be digitally recorded and transcribed ver-batim. Transcribed interviews and digital files will bepassword protected and stored on a password protectedcomputer in a secure, locked office. Digital files will bestored for 5 years and then deleted. All data will beanonymized for publication.

Analysis As recommended for qualitative descriptivestudies, we will use standard qualitative content analysisapproaches for thematic analysis of the transcripts [140].Two members of the team experienced in qualitativedata analysis will independently code the first two orthree transcripts, discuss, and reach consensus on a pre-liminary coding scheme. This coding scheme will be ap-plied in the coding of another two to three transcripts,following which the two researchers will discuss a re-vised coding scheme. At this point, the coding schemewill be sufficiently developed to allow one research teammember to independently code the remaining transcripts,with revisions made as necessary to reflect new and evolv-ing themes as data analysis progresses [140]. Thematicanalysis will occur concurrently with data collection toallow further exploration and clarification of emergentideas, and data collection will continue until data satur-ation is reached [141].


Trial statusThe IMPACT: Pilot and IMPACT: RCT trials were fundedwithin a few months of each other. Recruitment for theIMPACT: RCT has been deferred to February, 2014 topermit collection of pilot data. The pilot and full RCTshare the same study design and methodology, with theexception that the pilot study will collect data to 3 monthspostpartum, and the full RCT will collect data to 12months postpartum. The same recruitment sites will beutilized for both the pilot and full RCT. Data from thepilot trial will be used to refine the intervention, the eco-nomic evaluation, and the logistics involved in the fulltrial. At the time of manuscript submission, the CBTmodules and the online version of the psychosocial assess-ment component (ANRQ-R and EPDS) are being finalizedand recruitment for IMPACT: Pilot will begin in Septem-ber, 2013. Trial registration is through ClinicalTrials.gov(Identifier: NCT01901796).

AbbreviationsANRQ-R: Antenatal Risk Questionnaire; CBT: cognitive behavioral therapy;DASS21: Depression, Anxiety, and Stress Scale; EPDS: Edinburgh PostnatalDepression Scale; IMPACT: Integrated Maternal Psychosocial Assessment toCare Trial; QALY: Quality Adjusted Life Year; RCT: randomized controlled trial;SEM: structural equation modeling.

Competing interestsThe authors declare that they have no competing interests.

Authors’ contributionsDK conceived and designed the study, drafted the grant and the protocolmanuscript, organizes and supervises trial implementation, and is responsiblefor trial management, staff training and supervision. MPA, AB, KH, GL, SM,SDM, and SVZ participated in writing the grant. MPA, AB, KH, GL, WS, MH,SM, SDM, and SVZ contributed to the study design. KH, GL, SDM, TP, andSVZ participated in study implementation. MLS manages day-to-day trialresponsibilities, including supervising staff, monitoring recruitment and datacollection, and liaising with recruitment sites. SVZ provides expertise onstudy methodology and advises on trial management. MPA, KH, and GLprovide mental health expertise, AB, MH and SDM provide obstetrical expertise,and SM and RG provide child development expertise. SM provides statisticaland methodological expertise and DK, SM, and RG will conduct statisticalanalyses. MLS, DK, KH, and GL will conduct qualitative interviews and DK, GL,KH, MLS, MH and WS will analyze qualitative data. All authors contributed tothe development of the CBT modules. All authors participated in refinement ofthe study methods, critically reviewed manuscript drafts, and approved the finalmanuscript.

Authors’ informationDK (PhD) is an Assistant Professor in the Faculty of Nursing and an AdjunctAssistant Professor in the Department of Obstetrics and Gynecology at theUniversity of Alberta. She holds an Early Career Transition Award throughthe Alberta Centre for Child, Family, and Community Research. MPA(MD, FRANZCP, MB) is a perinatal psychiatrist and Professor in the Facultyof Medicine at University of New South Wales. She is also the Chair of thePerinatal and Women’s Mental Health Unit at the University of New SouthWales, the Director of the St. John of God Mother-Baby Unit in Sydney,Australia, and the lead developer of the Australian Clinical Guidelines forPerinatal Mental Health (2011) and the International Marce Society PositionStatement on Psychosocial Assessment and Depression Screening in the PerinatalPeriod (2013). AB (MD, CCFP, FCFP) is a family physician in the Mount SinaiAcademic Family Health Team and an Associate Professor in the Departmentof Family and Community Medicine at the University of Toronto. She holds theAda Slaight and Slaight Family Directorship in Maternity Care in the Ray D.Wolfe Department of Family Medicine at Mount Sinai Hospital in Toronto. RG(PhD) is a Senior Research Fellow and Clinical Psychologist at the Murdoch

Children's Research Institute, Melbourne, Australia. MH (PhD) is a Professor inthe Faculty of Nursing and an Associate Professor in the Departments ofCommunity Health Sciences and Obstetrics, Gynecology and ReproductiveSciences at the University of Manitoba. She held a CIHR Chair in Gender andHealth award from 2008-2013. KMH (PhD) is a Professor in the Faculty ofNursing and an Adjunct Professor in the Department of Psychiatry at theUniversity of Alberta. She holds a Canada Research Chair in Stress Disorders inWomen. GL (PhD) is an Associate Professor in the Faculty of Nursing at theUniversity of Alberta and is a Certified Psychiatric Nurse. SM (PhD) is anepidemiologist with expertise in life course analysis, mental health tooldevelopment, and child development. She is the senior scientist for the All OurBabies birth cohort study. SDM (MD, FRSCS, MSc) is an Associate Professor inthe Division of Maternal-Fetal Medicine in the Departments of Obstetrics andGynecology, Radiology, and Clinical Epidemiology and Biostatistics. Sheholds a CIHR New Investigator Award. TP (MD, CCFP, MSc) is a family physicianand an Assistant Professor in the Department of Family Medicine at McMasterUniversity. She is the Co-Head of Service in Family Medicine Obstetrics and theHamilton Site Coordinator of the Behaviour Science Program. WS (PhD) is aProfessor in the School of Nursing at McMaster University. MLS (MA) is theResearch Coordinator for the HOPE (Healthy Outcomes of Pregnancy andPostpartum Experiences) Program of research. SVZ (MD, PhD) is Director of theDivision of Gastroenterology at University of Alberta Hospital and a trialmethodologist.

AcknowledgementsThe authors would sincerely like to thank our recruitment sites for theirsupport. We would also like to thank Paper Leaf and AgileStyle for theirpartnership in developing the e-version of the CBT modules and ourstudents (Amy Toosi, Joshua Kingston, Karly Jarema, Lingfeng Zhu, and CaitlaneTarun) and staff (Michael Lee, Rhonda VanHeyst) for their contributions to thedevelopment of the CBT modules. We appreciate the informatics supportprovided by Rick Watts and Pamela Marples of the Women’s and Children’sHealth Research Institute (University of Alberta) in the construction and datamanagement of the e-questionnaire platform. Finally, it is with greatappreciation that Dr. Kingston thanks this research team for its strong support.The pilot trial is funded by the Norlien Foundation (http://www.norlien.org)and the Women’s and Children’s Health Research Institute (http://wchri.srv.ualberta.ca) and the full trial is funded by the Canadian Institutes of HealthResearch (CIHR) (http://www.cihr-irsc.gc.ca). The funders had no role in thedesign of the study and will not have a role in any other aspect of the trial,including its management, analysis or interpretation of data.

Author details1University of Alberta, 11405-87th Avenue, Edmonton T6G 1C9 AB, Canada.2University of New South Wales, Sydney, Australia. 3University of Calgary,Calgary, Canada. 4McMaster University, Hamilton, Canada. 5University ofManitoba, Winnipeg, Canada. 6University of Toronto, Toronto, Canada.7Murdoch Children’s Research Institute, Melbourne, Australia.

Received: 15 August 2013 Accepted: 13 February 2014Published: 6 March 2014

References1. Priest SR, Austin MP, Barnett BB, Buist A: A psychosocial risk assessment

model (PRAM) for use with pregnant and postpartum women in primarycare settings. Arch Womens Ment Health 2008, 11:307–317.

2. Grant KA, McMahon C, Austin MP: Maternal anxiety during the transitionto parenthood: a prospective study. J Affect Disord 2008, 108:101–111.

3. Austin MP, Hadzi-Pavlovic D, Priest SR, Reilly N, Wilhelm K, Saint K, Parker G:Depressive and anxiety disorders in the postpartum period: howprevalent are they and can we improve their detection? Arch WomensMent Health 2010, 13:395–401.

4. Austin MP, Tully L, Parker G: Examining the relationship between antenatalanxiety and postnatal depression. J Affect Disord 2007, 101:169–174.

5. Milgrom J, Gemmill AW, Bilszta JL, Hayes B, Barnett B, Brooks J, Ericksen J,Ellwood D, Buist A: Antenatal risk factors for postnatal depression: a largeprospective study. J Affect Disord 2008, 108:147–157.

6. Horwitz SM, Briggs-Gowan MJ, Storfer-Isser A, Carter AS: Persistence ofmaternal depressive symptoms throughout the early years of childhood.J Womens Health 2009, 18:637–645.

http://www.norlien.org

http://wchri.srv.ualberta.ca

http://wchri.srv.ualberta.ca

http://www.cihr-irsc.gc.ca


7. Sharpley C: Preliminary validation of the abbreviated Spanier DyadicAdjustment Scale. Educ Psychol Meas 1984, 44:1045–1050.

8. Mayberry LJ, Horowitz JA, Declercq E: Depression symptom prevalenceand demographic risk factors among U.S. women during the first 2 yearspostpartum. J Obstet Gynecol Neonatal Nurs 2007, 36:542–549.

9. Hobel CJ, Goldstein A, Barrett ES: Psychosocial stress and pregnancyoutcome. Clin Obstet Gynecol 2008, 51:333–348.

10. Beck CT: Maternal depression and child behaviour problems: ameta-analysis. J Adv Nurs 1999, 29:623–629.

11. Kingston D: The effects of prenatal and postpartum maternal psychologicaldistress on child development: a systematic review. The Centre for Child,Family and Community Research: Edmonton, Alberta; 2011.

12. Glasheen C, Richardson GA, Fabio A: A systematic review of the effects ofpostnatal maternal anxiety on children. Arch Womens Ment Health 2010,13:61–74.

13. Martini J, Knappe S, Beesdo-Baum K, Lieb R, Wittchen HU: Anxiety disordersbefore birth and self-perceived distress during pregnancy: associationswith maternal depression and obstetric, neonatal and early childhoodoutcomes. Early Hum Dev 2010, 86:305–310.

14. Howard LM, Flach C, Mehay A, Sharp D, Tylee A: The prevalence of suicidalideation identified by the Edinburgh Postnatal Depression Scale inpostpartum women in primary care: findings from the RESPOND trial.BMC Pregnancy Childbirth 2011, 11:57.

15. O’Hara MW, Wisner KL: Perinatal mental illness: Definition, descriptionand aetiology. Best Pract Res Clin Obstet Gynaecol 2014, 28:3–12.

16. Grigoriadis S, de Camps MD, Barrons E, Bradley L, Eady A, Fishell A,Mamisachvili L, Cook GS, O'Keefe M, Romans S, Ross LE: Mood and anxietydisorders in a sample of Canadian perinatal women referred forpsychiatric care. Arch Womens Ment Health 2011, 14:325–333.

17. Skouteris H, Wertheim EH, Rallis S, Milgrom J, Paxton SJ: Depression andanxiety through pregnancy and the early postpartum: an examination ofprospective relationships. J Affect Disord 2009, 113:303–308.

18. Kingston D, Tough S, Whitfield H: Prenatal and postpartum maternalpsychological distress and infant development: a systematic review.Child Psychiatry Hum Dev 2012, 43:683–714.

19. Shonkoff JP, Garner AS, Committee on Psychosocial Aspects of C, Family H,Committee on Early Childhood A, Dependent C, Section on D, Behavioral P:The lifelong effects of early childhood adversity and toxic stress.Pediatrics 2012, 129:e232–e246.

20. Coates AO, Schaefer CA, Alexander JL: Detection of postpartumdepression and anxiety in a large health plan. J Behav Health Serv Res2004, 31:117–133.

21. Spitzer RL, Williams JB, Kroenke K, Hornyak R, McMurray J: Validity andutility of the PRIME-MD patient health questionnaire in assessment of3000 obstetric-gynecologic patients: the PRIME-MD Patient HealthQuestionnaire Obstetrics-Gynecology Study. Am J Obstet Gynecol 2000,183:759–769.

22. Dennis CL, Chung-Lee L: Postpartum depression help-seeking barriersand maternal treatment preferences: a qualitative systematic review.Birth 2006, 33:323–331.

23. Flynn HA, Henshaw E, O'Mahen H, Forman J: Patient perspectives onimproving the depression referral processes in obstetrics settings: aqualitative study. Gen Hosp Psychiatry 2010, 32:9–16.

24. Reay R, Matthey S, Ellwood D, Scott M: Long-term outcomes ofparticipants in a perinatal depression early detection program. J AffectDisord 2011, 129:94–103.

25. Austin M-P, Highet N: Committee at GEA: Clinical practice guidelines fordepression and related disorders - anxiety, bipolar disorder and puerperalpsychosis - in the perinatal period. A guideline for primary care healthprofessionals. Committee at GEA: Melbourne, Australia; 2011.

26. Pignone MP, Gaynes BN, Rushton JL, Burchell CM, Orleans CT, Mulrow CD, LohrKN: Screening for depression in adults: a summary of the evidence for theU.S. Preventive Services Task Force. Ann Intern Med 2002, 136:765–776.

27. Buist A, Ellwood D, Brooks J, Milgrom J, Hayes BA, Sved-Williams A, BarnettB, Karatas J, Bilszta J: National program for depression associated withchildbirth: the Australian experience. Best Pract Res Clin Obstet Gynaecol2007, 21:193–206.

28. Chew-Graham C, Chamberlain E, Turner K, Folkes L, Caulfield L, Sharp D:GPs' and health visitors' views on the diagnosis and management ofpostnatal depression: a qualitative study. Br J Gen Pract 2008,58:169–176.

29. Leiferman JA, Dauber SE, Heisler K, Paulson JF: Primary care physicians'beliefs and practices toward maternal depression. J Womens Health 2008,17:1143–1150.

30. Reid AJ, Biringer A, Carroll JD, Midmer D, Wilson LM, Chalmers B, Stewart DE:Using the ALPHA form in practice to assess antenatal psychosocial health.Antenatal Psychosocial Health Assessment. CMAJ 1998, 159:677–684.

31. Austin MP, Colton J, Priest S, Reilly N, Hadzi-Pavlovic D: The Antenatal RiskQuestionnaire (ANRQ-R): Acceptability and use for psychosocial riskassessment in the maternity setting. Women Birth 2013, 26:17–25.

32. Buist A, Condon J, Brooks J, Speelman C, Milgrom J, Hayes B, Ellwood D,Barnett B, Kowalenko N, Matthey S, Austin MP, Bilszta J: Acceptability ofroutine screening for perinatal depression. J Affect Disord 2006, 93:233–237.

33. Matthey S, White T, Phillips J, Taouk R, Chee TT, Barnett B: Acceptability ofroutine antenatal psychosocial assessments to women from English and non-English speaking backgrounds. Arch Womens Ment Health 2005, 8:171–180.

34. Coleman VH, Carter MM, Morgan MA, Schulkin J: Obstetrician-gynecologists'screening patterns for anxiety during pregnancy. Depress Anxiety 2008,25:114–123.

35. Kim JJ, La Porte LM, Corcoran M, Magasi S, Batza J, Silver RK: Barriers tomental health treatment among obstetric patients at risk for depression.Am J Obstet Gynecol 2010, 202:312–315.

36. Bowen A, Bowen R, Butt P, Rahman K, Muhajarine N: Patterns ofdepression and treatment in pregnant and postpartum women. Can JPsychiatry 2012, 57:161–167.

37. Sword W, Busser D, Ganann R, McMillan T, Swinton M: Women's care-seeking experiences after referral for postpartum depression. Qual HealthRes 2008, 18:1161–1173.

38. Woolhouse H, Brown S, Krastev A, Perlen S, Gunn J: Seeking help foranxiety and depression after childbirth: results of the Maternal HealthStudy. Arch Womens Ment Health 2009, 12:75–83.

39. Chew-Graham CA, Sharp D, Chamberlain E, Folkes L, Turner KM: Disclosureof symptoms of postnatal depression, the perspectives of healthprofessionals and women: a qualitative study. BMC Fam Pract 2009, 10:7.

40. Miller L, Shade M, Vasireddy V: Beyond screening: assessment of perinataldepression in a perinatal care setting. Arch Womens Ment Health 2009, 12:329–334.

41. Butler M, Kane R, McAlpine D, Kathol R, Fu S, Hagedorn H, Wilt T: Rockville,MD: Agency for Healthcare Research and Quality (AHRQ). In Integration ofMental Health/Substance Abuse and Primary Care: Evidence Report/TechnologyAssessment. 2008.

42. Bartels SJ, Coakley EH, Zubritsky C, Ware JH, Miles KM, Arean PA, Chen H, OslinDW, Llorente MD, Costantino G, Quijano L, McIntyre JS, Linkins KW, Oxman TE,Maxwell J, Levkoff SE, PRISM-E Investigators: Improving access to geriatricmental health services: a randomized trial comparing treatment engagementwith integrated versus enhanced referral care for depression, anxiety, andat-risk alcohol use. Am J Psychiatry 2004, 161:1455–1462.

43. Butler M, Kane RL: Does integrated care improve treatment fordepression? A systematic review. J Ambul Care Manage 2011, 34:113–125.

44. Neumeyer-Gromen A, Lampert T, Stark K, Kallischnigg G: Diseasemanagement programs for depression: a systematic review andmeta-analysis of randomized controlled trials. Med Care 2004,42:1211–1221.

45. Flynn HA, O'Mahen HA, Massey L, Marcus S: The impact of a briefobstetrics clinic-based intervention on treatment use for perinataldepression. J Womens Health 2006, 15:1195–1204.

46. Gjerdingen D, Katon W, Rich DE: Stepped care treatment of postpartumdepression: a primary care-based management model. Womens HealthIssues 2008, 18:44–52.

47. Horwitz SM, Kelleher KJ, Stein RE, Storfer-Isser A, Youngstrom EA, Park ER,Heneghan AM, Jensen PS, O'Connor KG, Hoagwood KE: Barriers to theidentification and management of psychosocial issues in children andmaternal depression. Pediatrics 2007, 119:e208–e218.

48. Leddy M, Haaga D, Gray J, Schulkin J: Postpartum mental health screeningand diagnosis by obstetrician-gynecologists. J Psychosom Obstet Gynaecol2011, 32:27–34.

49. Gordon TE, Cardone IA, Kim JJ, Gordon SM, Silver RK: Universal perinataldepression screening in an Academic Medical Center. Obstet Gynecol2006, 107:342–347.

50. Hewitt C, Gilbody S, Brealey S, Paulden M, Palmer S, Mann R, Green J,Morrell J, Barkham M, Light K, Richards D: Methods to identify postnataldepression in primary care: an integrated evidence synthesis and valueof information analysis. Health Technol Assess 2009, 13:1–230.


51. Carroll JC, Reid AJ, Biringer A, Midmer D, Glazier RH, Wilson L, Permaul JA,Pugh P, Chalmers B, Seddon F, Stewart DE: Effectiveness of the AntenatalPsychosocial Health Assessment (ALPHA) form in detecting psychosocialconcerns: a randomized controlled trial. CMAJ 2005, 173:253–259.

52. Mitchell AJ, Coyne J: Screening for postnatal depression: barriers tosuccess. BJOG 2009, 116:11–14.

53. Renker PR, Tonkin P: Women's views of prenatal violence screening:acceptability and confidentiality issues. Obstet Gynecol 2006, 107:348–354.

54. Renker PR, Tonkin P: Postpartum women's evaluations of an audio/videocomputer-assisted perinatal violence screen. Comput Inform Nurs 2007,25:139–147.

55. Le HN, Perry DF, Sheng X: Using the internet to screen for postpartumdepression. Matern Child Health J 2009, 13:213–221.

56. Renker PR: Breaking the barriers: the promise of computer-assistedscreening for intimate partner violence. J Midwifery Womens Health 2008,53:496–503.

57. Choo EK, Ranney ML, Aggarwal N, Boudreaux ED: A systematic review ofemergency department technology-based behavioral health interventions.Acad Emerg Med 2012, 19:318–328.

58. Locke SE, Kowaloff HB, Hoff RG, Safran C, Popovsky MA, Cotton DJ,Finkelstein DM, Page PL, Slack WV: Computer interview for screeningblood donors for risk of HIV transmission. MD Comput 1994, 11:26–32.

59. Turner CF, Ku L, Rogers SM, Lindberg LD, Pleck JH, Sonenstein FL:Adolescent sexual behavior, drug use, and violence: increased reportingwith computer survey technology. Science 1998, 280:867–873.

60. Gilbody S, Whitty P, Grimshaw J, Thomas R: Educational and organizationalinterventions to improve the management of depression in primarycare: a systematic review. JAMA 2003, 289:3145–3151.

61. Rowan P, Greisinger A, Brehm B, Smith F, McReynolds E: Outcomes fromimplementing systematic antepartum depression screening in obstetrics.Arch Womens Ment Health 2012, 15:115–120.

62. Cape J, Whittington C, Buszewicz M, Wallace P, Underwood L: Briefpsychological therapies for anxiety and depression in primary care:meta-analysis and meta-regression. BMC Med 2010, 8:38.

63. Tolin D: Is cognitive-behavioral therapy more effective than othertherapies? A meta-analytic review. Clin Psychol Rev 2010, 30:710–720.

64. Chabrol H: CBT versus supportive therapy for depression. J Am Acad ChildAdolesc Psychiatry 2005, 44:841–843.

65. Goodman JH, Santangelo G: Group treatment for postpartum depression:a systematic review. Arch Womens Ment Health 2011, 14:277–293.

66. Milgrom J, Negri LM, Gemmill AW, McNeil M, Martin PR: A randomizedcontrolled trial of psychological interventions for postnatal depression.Br J Clin Psychol 2005, 44:529–542.

67. Sockol LE, Epperson CN, Barber JP: A meta-analysis of treatments forperinatal depression. Clin Psychol Rev 2011, 31:839–849.

68. Stevenson MD, Scope A, Sutcliffe PA, Booth A, Slade P, Parry G, Saxon D,Kalthenthaler E: Group cognitive behavioural therapy for postnataldepression: a systematic review of clinical effectiveness, cost-effectivenessand value of information analyses. Health Technol Assess 2010, 14:1–iv.

69. Austin MP, Frilingos M, Lumley J, Hadzi-Pavlovic D, Roncolato W, Acland S,Saint K, Segal N, Parker G: Brief antenatal cognitive behaviour therapy groupintervention for the prevention of postnatal depression and anxiety:a randomised controlled trial. J Affect Disord 2008, 105:35–44.

70. Cho HJ, Kwon JH, Lee JJ: Antenatal cognitive-behavioral therapy forprevention of postpartum depression: a pilot study. Yonsei Med J 2008,49:553–562.

71. Proudfoot J: The development and beta-test of a computer-therapyprogram for anxiety and depression: hurdles and lessons. Comput HumanBehav 2003, 19:277–289.

72. Austin MP: Targeted group antenatal prevention of postnatal depression:a review. Acta Psychiatr Scand 2003, 107:244–250.

73. Horowitz JA, Goodman J: A longitudinal study of maternal postpartumdepression symptoms. Res Theory Nurs Pract 2004, 18:149–163.

74. Horwitz SM, Briggs-Gowan MJ, Storfer-Isser A, Carter AS: Prevalence, correlates,and persistence of maternal depression. J Womens Health 2007, 16:678–691.

75. Petrou S, Cooper P, Murray L, Davidson LL: Economic costs of post-natal depression in a high-risk British cohort. Br J Psychiatry 2002,181:505–512.

76. El-Mohandes AA, Kiely M, Gantz MG, El-Khorazaty MN: Very preterm birth isreduced in women receiving an integrated behavioral intervention: arandomized controlled trial. Matern Child Health J 2011, 15:19–28.

77. Kozinszky Z, Dudas RB, Devosa I, Csatordai S, Toth E, Szabo D, Sikovanyecz J,Barabas K, Pal A: Can a Brief Antepartum Preventive Group InterventionHelp Reduce Postpartum Depressive Symptomatology? PsychotherPsychosom 2012, 81:98–107.

78. Chaput K, Vekved M, McDonald S, McGrath P, Tough S: Pilot testing aprenatal telephone-based intervention for the prevention of postpartumdepression. Can J Public Health 2012.

79. Clark DM, Layard R, Smithies R, Richards DA, Suckling R, Wright B:Improving access to psychological therapy: Initial evaluation of two UKdemonstration sites. Behav Res Ther 2009, 47:910–920.

80. Proudfoot J, Goldberg D, Mann A, Everitt B, Marks I, Gray JA: Computerized,interactive, multimedia cognitive-behavioural program for anxiety anddepression in general practice. Psychol Med 2003, 33:217–227.

81. Proudfoot J, Ryden C, Everitt B, Shapiro DA, Goldberg D, Mann A, Tylee A,Marks I, Gray JA: Clinical efficacy of computerised cognitive-behaviouraltherapy for anxiety and depression in primary care: randomisedcontrolled trial. Br J Psychiatry 2004, 185:46–54.

82. Proudfoot JG: Computer-based treatment for anxiety and depression: is itfeasible? Is it effective? Neurosci Bio Behav Rev 2004, 28:353–363.

83. NICE: Computerised cognitive behavior therapy for depression and anxiety.United Kingdom: National Institute for Health and Clinical Excellence; 2008.

84. Andrews G: Computer therapy for the anxiety and depressive disordersis effective, acceptable and practical health care: a meta-analysis. PLoSOne 2010, 5:e13196.

85. Ugarriza DN: Group therapy and its barriers for women suffering frompostpartum depression. Arch Psychiatr Nurs 2004, 18:39–48.

86. Rini CK, Dunkel-Schetter C, Wadhwa PD, Sandman CA: Psychologicaladaptation and birth outcomes: the role of personal resources,stress, and sociocultural context in pregnancy. Health Psychol 1999,18:333–345.

87. Farrer L, Christensen H, Griffiths KM, Mackinnon A: Internet-based CBT fordepression with and without telephone tracking in a national helpline:randomised controlled trial. PLoS One 2011, 6:e28099.

88. Berger T, Hammerli K, Gubser N, Andersson G, Caspar F: Internet-basedtreatment of depression: a randomized controlled trial comparingguided with unguided self-help. Cogn Behav Ther 2011, 40:251–266.

89. Hewitt CE, Gilbody SM: Is it clinically and cost effective to screen forpostnatal depression: a systematic review of controlled clinical trials andeconomic evidence. BJOG 2009, 116:1019–1027.

90. Campbell NC, Murray E, Darbyshire J, Emery J, Farmer A, Griffiths F, GuthrieB, Lester H, Wilson P, Kinmonth AL: Designing and evaluating complexinterventions to improve health care. BMJ 2007, 334:455–459.

91. Connor KaD J: Development of a new resilience scale: The Connor-Davidson Resilience Scale (CD-RISC). Depress Anxiety 2003, 18:76–82.

92. Bale TL, Baram TZ, Brown AS, Goldstein JM, Insel TR, McCarthy MM, NemeroffCB, Reyes TM, Simerly RB, Susser ES, Nestler EJ: Early life programming andneurodevelopmental disorders. Biol Psychiatry 2010, 68:314–319.

93. Robinson M, Mattes E, Oddy WH, Pennell CE, van Eekelen A, McLean NJ,Jacoby P, Li J, De Klerk NH, Zubrick SR, Stanley FJ, Newnham JP: Prenatalstress and risk of behavioral morbidity from age 2 to 14 years: Theinfluence of the number, type, and timing of stressful life events. DevPsychopathol 2011, 23:507–520.

94. Pemberton CK, Neiderhiser JM, Leve LD, Natsuaki MN, Shaw DS, Reiss D, GeX: Influence of parental depressive symptoms on adopted toddlerbehaviors: an emerging developmental cascade of genetic andenvironmental effects. Dev Psychopathol 2010, 22:803–818.

95. Buss C, Davis EP, Muftuler LT, Head K, Sandman CA: High pregnancy anxietyduring mid-gestation is associated with decreased gray matter density in6-9-year-old children. Psychoneuroendocrinology 2010, 35:141–153.

96. Davis EP, Glynn LM, Waffarn F, Sandman CA: Prenatal maternal stressprograms infant stress regulation. J Child Psychol Psychiatry 2011,52:119–129.

97. Sandman CA, Davis EP, Buss C, Glynn LM: Exposure to prenatalpsychobiological stress exerts programming influences on the motherand her fetus. Neuroendocrinology 2012, 95:7–21.

98. Urizar GG Jr, Munoz RF: Impact of a prenatal cognitive-behavioral stressmanagement intervention on salivary cortisol levels in low-income mothersand their infants. Psychoneuroendocrinology 2011, 36:1480–1494.

99. Austin MP, Priest SR, Sullivan EA: Antenatal psychosocial assessment forreducing perinatal mental health morbidity. Cochrane Database Syst Rev2008, 4, CD005124.


100. Petrou S: Economic consequences of preterm birth and low birthweight.BJOG 2003, 110(Suppl 20):17–23.

101. Petrou S: The economic consequences of preterm birth during the first10 years of life. BJOG 2005, 112(Suppl 1):10–15.

102. Petrou S, Mehta Z, Hockley C, Cook-Mozaffari P, Henderson J, Goldacre M:The impact of preterm birth on hospital inpatient admissions and costsduring the first 5 years of life. Pediatrics 2003, 112:1290–1297.

103. Waddell C, McEwan K, Shepherd CA, Offord DR, Hua JM: A public healthstrategy to improve the mental health of Canadian children. Can JPsychiatry 2005, 50:226–233.

104. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M:Developing and evaluating complex interventions: the new MedicalResearch Council guidance. BMJ 2008, 337:a1655.

105. Hulley SB, Cummings SR, Browner WS, Grady DG, Newman TB: Designing ClinicalResearch. 3rd edition. Philadelphia, PA: Lippincott Williams & Wilkins; 2007.

106. Leigh B, Milgrom J: Acceptability of antenatal screening for depression inroutine antenatal care. Aust J Adv Nurs 2007, 24:14–18.

107. Lovibond SH, Lovibond PF: Manual for the Depression Anxiety Stress Scales.Sydney, Australia: Psychology Foundation; 1995.

108. Ftanou M, Bassilios B, Fletcher J, King K, Kohn F, Blashki G, Burgess P, Pirkis J:The Third Report of the Perinatal Depression Initiative. In Evaluating theAccess to Allied Psychological Services. Melbourne, AU: Centre for HealthPolicy, Programs, and Economics; 2011.

109. Kline RB: Principles and Practice of Structural Equation Modeline. 3rd edition.New York, N.Y.: The Guilford Press; 2011.

110. Milgrom J, Schembri C, Ericksen J, Ross J, Gemmill AW: Towardsparenthood: an antenatal intervention to reduce depression, anxiety andparenting difficulties. J Affect Disord 2011, 130:385–394.

111. Cox JL, Holden JM, Sagovsky R: Detection of postnatal depression.Development of the 10-item Edinburgh Postnatal Depression Scale. Br JPsychiatry 1987, 150:782–786.

112. Flynn HA, Sexton M, Ratliff S, Porter K, Zivin K: Comparative performanceof the Edinburgh Postnatal Depression Scale and the Patient HealthQuestionnaire-9 in pregnant and postpartum women seeking psychiatricservices. Psychiatry Res 2011, 187:130–134.

113. Barkham M, Shapiro DA, Hardy GE, Rees A: Psychotherapy in two-plus-onesessions: outcomes of a randomized controlled trial of cognitive-behavioral and psychodynamic-interpersonal therapy for subsyndromaldepression. J Consult Clin Psychol 1999, 67:201–211.

114. Scott C, Tacchi MJ, Jones R, Scott J: Acute and one-year outcome of arandomised controlled trial of brief cognitive therapy for majordepressive disorder in primary care. Br J Psychiatry 1997, 171:131–134.

115. Osman A, Wong JL, Bagge CL, Freedenthal S, Gutierrez PM, Lozano G: TheDepression Anxiety Stress Scales-21 (DASS-21): further examination ofdimensions, scale reliability, and correlates. J Clin Psychology 2012,68:1322–1338.

116. Sinclair S: Psychometric Evaluation and Mormative Data for the Depression,Anxiety, and Stress Scales-21 (DASS-21) in a Nonclinical Sample of U.S.Adults. Eval Health Prof 2012, 35:259–279.

117. de Paz NC, Sanchez SE, Huaman LE, Chang GD, Pacora PN, Garcia PJ, AnanthCV, Qiu C, Williams MA: Risk of placental abruption in relation to maternaldepressive, anxiety and stress symptoms. J Affect Disord 2011, 130:280–284.

118. Antony M: Psychometric properties of the 42-item and 21-item versionsof the depression, anxiety, stress scales (DASS) in clinical groups and acommunity sample. Psychol Assess 1998, 10:176–181.

119. Brown TA, Chorpita BF, Korotitsch W, Barlow DH: Psychometric propertiesof the Depression Anxiety Stress Scales (DASS) in clinical samples. BehavRes Ther 1997, 35:79–89.

120. Dzakpasu S, Kaczorowski J, Chalmers B, Heaman M, Duggan J, Neusy E,Maternity Experiences Study Group of the Canadian Perinatal SurveillanceSystem, Public Health Agency of Canada: The Canadian maternityexperiences survey: design and methods. JOGC 2008, 30:207–216.

121. Cohen S, Mermelstein R, Kamarck T, Hoberman HM: Measuring thefunctional components of social support. In Social Support: Theory,Research, and Applications. Martinus Nijhoff Publishers: Sarason IG, SarasonBR. Dordrecht; 1985:73–94.

122. Pearlin LI, Schooler C: The structure of coping. J Health Soc Behav 1978,19:2–21.

123. Schwarzer R, Jerusalem M: Generalized self-efficacy scale. In Measures inhealth psychology: a user's portfolio causal and control beliefs. Windsor, U.K:NFER-NELSON; 1995:35–37.

124. Rosenberg M: Society and the adolescent self-image. Princeton, N.J.: PrincetonUniversity Press; 1965.

125. Backhaus J, Junghanns K, Broocks A, Riemann D, Hohagen F: Test-retestreliability and validity of the Pittsburgh Sleep Quality Index in primaryinsomnia. J Psychosom Res 2002, 53:737–740.

126. Skouteris H, Wertheim EH, Germano C, Paxton SJ, Milgrom J: Assessingsleep during pregnancy: a study across two time points examining thePittsburgh Sleep Quality Index and associations with depressivesymptoms. Womens Health Issues 2009, 19:45–51.

127. Berry JO, Jones WH: The Parental Stress Scale: Initial psychometricevidence. J Soc Pers Relat 1995, 12:463–472.

128. Gilmore L, Cuskelly M: Factor structure of the Parenting Sense ofCompetence scale using a normative sample. Child Care Health Dev 2009,35:48–55.

129. Whisman MA, Davila J, Goodman SH: Relationship adjustment, depression,and anxiety during pregnancy and the postpartum period. J Fam Psychol2011, 25:375–383.

130. Carver CS, Scheier MF, Weintraub JK: Assessing coping strategies: atheoretically based approach. J Pers Soc Psychol 1989, 56:267–283.

131. Condon J, Corkindale C: The assessment of parent-to-infant attachment:Development of a self-report questionnaire instrument. J Reprod InfantPsychol 1998, 16:19.

132. Gartstein MA, Rothbart MK: Studying infant temperament via the RevisedInfant Behavior Questionnaire. Infant Behav Dev 2003, 26:64–86.

133. Sheldrick RC, Henson BS, Neger EN, Merchant S, Murphy JM, Perrin EC: Thebaby pediatric symptom checklist: development and initial validation ofa new social/emotional screening instrument for very young children.Acad Pediatr 2013, 13:72–80.

134. Koopmanschap M: PRODISQ: a modular questionnaire on productivityand disease for economic evaluation studies. Expert Rev PharmacoeconOutcomes Res 2005, 5:23–28.

135. Stewart AK: Measuring Functioning and Well-Being: The Medical OutcomesStudy Approach. Durham, N.C.: Duke University Press; 1992.

136. McDonald SW, Lyon AW, Benzies KM, McNeil DA, Lye SJ, Dolan SM, Pennell CE,Bocking AD, Tough SC: The All Our Babies pregnancy cohort: design,methods, and participant characteristics. BMC Pregnancy Childbirth 2013, 13:1.

137. Dillman D: Mail and Internet Surveys. 2nd edition. Hoboken, New Jersey:John Wiley & Sons, Inc; 2007.

138. Gold M: Panel on cost-effectiveness in health and medicine. Med Care1996, 34(12 Suppl):DS197–DS199.

139. Candy B, King M, Jones L, Oliver S: Using qualitative evidence on patients'views to help understand variation in effectiveness of complexinterventions: A qualitative comparative analysis. BMC: Trials 2013, 14:179.

140. Sandelowski M: Whatever happened to qualitative description? Res NursHealth 2000, 23:334–340.

141. Manning P: Narrative, content, and semiotic analysis. In Handbook ofQualitative Research. Edited by Denzin N. L., Y. London: Sage; 1994.

doi:10.1186/1745-6215-15-72Cite this article as: Kingston et al.: Study protocol for a randomized,controlled, superiority trial comparing the clinical andcost- effectiveness of integrated online mental health assessment-referral-care in pregnancy to usual prenatal care on prenatal andpostnatal mental health and infant health and development: theIntegrated Maternal Psychosocial Assessment to Care Trial (IMPACT). Trials2014 15:72.

study protocol open access study protocol for a randomized, … · 2017-08-25 · cognitive...

Documents